Some thoughts on the decline of unemployment in Kerala

R. 

In the last post on this blog, we had looked at some data on the fall in unemployment rates in Kerala between 2004-05 and 2009-10. In this post, i would like to discuss some of the reasons as to why scholars thought unemployment rates were high in Kerala till 2004-05. Till then, Kerala had one of the highest unemployment rates in India. Let me make use of two important writings on the subject.

1) Mridul Eapen (1992), “Fertility and Female Labour Force Participation In Kerala”, Economic and Political Weekly, 27 (40), October 3, pp. 2179-2188.

2) T. M. Thomas Isaac (2000), “Women and Work”, in Studies on Status of Women, State Planning Board, July, pp.21-74 (in Malayalam).

First, it was argued that labour force participation rates (LFPRs) in Kerala are reduced by high levels of unemployment through a discouraged worker effect, especially for females. For beginners, LFPR denotes the share of the labour force (i.e., workers plus the unemployed; "unemployed" is defined as those looking for work, but could not find one) to the population, expressed in percentage. Work participation rate (WPR) is defined as the percentage of total workers to the total population. Typically, a high rate of unemployment discourages people who search for jobs through a rise in the job search costs. High job search costs eventually push people out of the "labour force" as they stop looking for jobs. This phenomenon is referred to in the labour economics literature as discouraged worker effect. Discouraged worker effect is thus defined as a situation where people who want a job, and are currently available for work, give up active job search because they believe they cannot find a job. 

Mridul Eapen argued that “high rates of unemployment have discouraged female entrants [into the labour force]…forcing them into economic inactivity or into marginal work or the type of work which escapes census enumeration”. Thomas Isaac argued that “unemployment rate among young women is high…Even after long waiting periods, majority of young women remain unemployed. Hence, they cease to be job seekers and confine themselves to their houses. As they cease to be job seekers, they are classified as to be out of the labour force by official surveys”

One way to check for discouragement in work seeking from secondary data is to look at the relationship between economy-wide unemployment rates and LFPRs. LFPR, unlike WPR, includes those who are currently working and those who are actively looking for work. If there is an inverse relationship between unemployment rates and LFPRs over time, then it gives an indication that increasing unemployment is associated with more number of people ceasing to search for jobs. It would lend support to the argument that absence of jobs in the locality is forcing work seekers to opt out of the labour force. An analysis that I tried with NSS data for rural Kerala showed that such an inverse relationship was evident for women, but not for men: that is, in the long run, between 1972-73 and 1999-00, as the unemployment rate went up in the economy, female LFPRs declined. 

The second important reason noted for low work participation rates is the lack of employment opportunities that match the educational achievements of a job seeker. A large proportion of unemployed persons in Kerala is educated, due to the spread of education across classes and castes in the post-1957 period. According to one estimate, 16.1 per cent of the educated unemployed in India belonged to Kerala in 1987-88. The most important effect of the spread of education on the labour market was the transformation of job expectations of the educated job seekers. As Chandan Mukherji and Thomas Isaac have noted in a study in 1991, the spread of education in Kerala created a definite preference for jobs in the formal service sector (see Chandan Mukherjee and T. M. Thomas Isaac (1991), “Study of Educated Unemployed in Kerala: Report of the Sample Survey of Restraints of Employment Exchange”, Planning Commission, New Delhi and Centre for Development Studies, Thiruvananthapuram). The restricted number of jobs available in the formal sector led to a fall in WPRs, especially for women.

Now, what appears to have happened in Kerala after 2004-05 is a reversal of both the above trends noted for earlier years. After 2004-05, the number of jobs created has risen and these new jobs appear to be those that were more compatible with the aspirations of the educated unemployed. The fact that both the labour force and the work force have expanded show that more people have entered the labour force, and not just that, they were able to get jobs and thus, enter the work force also. So, the impact of the discouraged worker effect has also diminished in the last 5-6 years. This was particularly so for women in both rural and urban areas.

How did this happen? We can have some preliminary arguments. There are three things that has happened in the last 5 years on the employment front in Kerala. And none of us can decompose it realistically to separate them and find individual effects.

First, there has been a major effort to protect employment in sectors where there historically was loss of employment, and thus a rise in unemployment rates. The two important examples here are the LDF's govt's efforts to protect employment in (a) the traditional industrial sector and (b) the public sector industries. These efforts made sure that employment does not fall in these sectors where it has always shown a tendency to fall. Let us call it a preventive step.

Secondly, there was an effort to raise employment through effective implementation of NREGS in specific regions and Kudumbashree. Both these are likely to have made a major impact on female employment. They are likely to have reduced the "discouraged worker effect" significantly and brought women in large numbers from outside the labour force into the labour force. This explains the rise in LFPRs and fall in unemployment rates. Let us call this a promotional step.

Thirdly, there is likely to have been an independent impact on jobs due to general economic growth. If we decompose economic growth after 2005, we see that growth has come from industries (manufacturing) and services (construction, transport, storage and communication, trade, hotels and restaurants, banking and insurance and real estate). This is a result of growth domestically and not due to remittances. There is enough evidence to show that it was only during two years (2008-09 and 2009-10) that remittances rose. In these 2 years, bank deposits have risen sharply, but it was largely because Non-Resident Keralites (NRKs) shifted deposits from foreign banks (that were crashing) into public sector domestic banks in India. It has already ended, and in 2010-11, deposits have already fallen. So, remittances have little to do with this. This form of growth is likely to have given rise to regular jobs, and it is here that the impact on educated unemployment is likely to have been highest.

There is some data to prove this point also. See the following table, that gives the share of employment by three categories, 1) self-employed (SE); 2) regular employment; and 3) casual employment, for 2004-05 and 2009-10.

Distribution of usually employed (principal+subsidiary) 

by category of employment, Kerala,2004-05 and 2009-10, in %

Source: NSS reports.

Here, we can clearly see that the share of regular employment has risen from 16% to 19% in rural areas and from 28.7% to 34.2% in urban areas. This is a sharp rise, and all this is likely to be educated/skilled employment. Look at the pattern that shows that this rise for women is more than for men. In particular, urban women (38% in 2004-05 to 48% in 2009-10). On the other hand, other processes at work (including NREGS, Kudumbashree and informal sector growth associated with growth in general) are likely to have contributed to the increase in the share of casual employment. This has risen for rural women and urban men. 

What happened in India during the same period? In India as a whole, and in rural and urban areas, what has happened is that the work force itself has shrunk. Between 2004-05 and 2009-10, the worker-population ratio in India as a whole fell from 38 per cent to 36.5 per cent. In rural India, it fell from 39.1 per cent to 37.4 per cent. In urban India, it fell from 34.6 per cent to 33.9 per cent. On the other hand, the worker-population ratio in Kerala (as shown in Table 2 of my last post) has increased between 2004-05 and 2009-10. This is the most important difference between India and Kerala, even though you see a small fall in unemployment rate for India as well.

So, the work force has not expanded in India at all in this period. It has expanded in Kerala; new people have entered the work force. Labour force (work force+unemployed) has also expanded, but work force has grown faster than labour force in Kerala. For those interested in the all-India picture, here is Professor C. P. Chandrasekhar's blog entry in The Hindu website:

http://www.thehindu.com/opinion/columns/Chandrasekhar/article2153082.ece

Does it mean that Kerala is entering a new phase in its development? Is Kerala's jobless growth a history? What really gets questioned here is a basic assumption in Kerala studies till now: that the growth effects of Kerala spill out to other States and thus, jobs are not domestically created. This appears to be passe! A particular focus on the nature of growth can make sure that jobs are domestically created. The second half of the 2000s in Kerala appears to be a success story on that front. Only more intensive studies can unravel this phenomenon fully.